{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:FW6K2W2XWJUVXIE7X3FNH5TNZ5","short_pith_number":"pith:FW6K2W2X","schema_version":"1.0","canonical_sha256":"2dbcad5b57b2695ba09fbecad3f66dcf477412d527aff732d12cd14bdf5c3700","source":{"kind":"arxiv","id":"2411.17866","version":2},"attestation_state":"computed","paper":{"title":"Distributed Sign Momentum with Local Steps for Training Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"An Xu, Cong Xie, Ding Zhou, Shuhua Yu, Soummya Kar, Xin Liu, Zhi Zhang","submitted_at":"2024-11-26T20:31:11Z","abstract_excerpt":"Pre-training Transformer models is resource-intensive, and recent studies have shown that sign momentum is an efficient technique for training large-scale deep learning models, particularly Transformers. However, its application in distributed training remains underexplored. This paper investigates a novel communication-efficient distributed sign momentum method with multiple local steps, to cope with the scenarios where communicating at every step is prohibitive. Our proposed method allows for a broad class of base optimizers for local steps, and uses sign momentum in the global step, where m"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2411.17866","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-11-26T20:31:11Z","cross_cats_sorted":[],"title_canon_sha256":"65204abf8f27e5e8bf228e2f65eb6180986e153eb29678972f4ec2f2b3725766","abstract_canon_sha256":"0891f838c80b548ffc50cac6b8b450a943197a90b2efefa7c5c7bb5f22e72e61"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:26:41.738028Z","signature_b64":"3nPjFXrsXcJ3Z9V4ur0h5fzBlPX934DnH4nmc3x30Lck9SFuCxIKWORqfl+BTc+8bl3+lXrnUqTuZHcEZZghBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2dbcad5b57b2695ba09fbecad3f66dcf477412d527aff732d12cd14bdf5c3700","last_reissued_at":"2026-07-05T10:26:41.737409Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:26:41.737409Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Distributed Sign Momentum with Local Steps for Training Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"An Xu, Cong Xie, Ding Zhou, Shuhua Yu, Soummya Kar, Xin Liu, Zhi Zhang","submitted_at":"2024-11-26T20:31:11Z","abstract_excerpt":"Pre-training Transformer models is resource-intensive, and recent studies have shown that sign momentum is an efficient technique for training large-scale deep learning models, particularly Transformers. However, its application in distributed training remains underexplored. This paper investigates a novel communication-efficient distributed sign momentum method with multiple local steps, to cope with the scenarios where communicating at every step is prohibitive. Our proposed method allows for a broad class of base optimizers for local steps, and uses sign momentum in the global step, where m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.17866","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2411.17866/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2411.17866","created_at":"2026-07-05T10:26:41.737481+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.17866v2","created_at":"2026-07-05T10:26:41.737481+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.17866","created_at":"2026-07-05T10:26:41.737481+00:00"},{"alias_kind":"pith_short_12","alias_value":"FW6K2W2XWJUV","created_at":"2026-07-05T10:26:41.737481+00:00"},{"alias_kind":"pith_short_16","alias_value":"FW6K2W2XWJUVXIE7","created_at":"2026-07-05T10:26:41.737481+00:00"},{"alias_kind":"pith_short_8","alias_value":"FW6K2W2X","created_at":"2026-07-05T10:26:41.737481+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5","json":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5.json","graph_json":"https://pith.science/api/pith-number/FW6K2W2XWJUVXIE7X3FNH5TNZ5/graph.json","events_json":"https://pith.science/api/pith-number/FW6K2W2XWJUVXIE7X3FNH5TNZ5/events.json","paper":"https://pith.science/paper/FW6K2W2X"},"agent_actions":{"view_html":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5","download_json":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5.json","view_paper":"https://pith.science/paper/FW6K2W2X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.17866&json=true","fetch_graph":"https://pith.science/api/pith-number/FW6K2W2XWJUVXIE7X3FNH5TNZ5/graph.json","fetch_events":"https://pith.science/api/pith-number/FW6K2W2XWJUVXIE7X3FNH5TNZ5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5/action/storage_attestation","attest_author":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5/action/author_attestation","sign_citation":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5/action/citation_signature","submit_replication":"https://pith.science/pith/FW6K2W2XWJUVXIE7X3FNH5TNZ5/action/replication_record"}},"created_at":"2026-07-05T10:26:41.737481+00:00","updated_at":"2026-07-05T10:26:41.737481+00:00"}